查看更多>>摘要：Service quality is significant for improving customer satisfaction and enhancing competitive advantages of service business. Due to the dynamic nature of service activities, system simulation has become popular for researching service operations management during the past few years. However, most previous studies about service operations system simulation use only one single modeling paradigm, making it still challenging for exploring individual consumers' behaviors and evaluating the relative efficiency of different alternative strategies for service operations improvement. In this context, a hybrid modeling methodology framework is proposed in this study for integrating customers' behaviors into service operations system simulation. Specifically, we designed an algorithm for simulating consumers' behavioral decisions based on cognitive psychology. And then, individual consumer's decision behavior was modeled as an agent, and integrated into a hybrid simulation model developed based on DES (Discrete Event Simulation) coupled with ABS (Agent-Based Simulation) paradigms. Moreover, we involve cost-benefit analysis for further exploring different alternative operation improvement strategies based on simulation experiments. We then design and implement a prototype simulator based on the methodology framework and apply it in a real case study to demonstrate its functionalities and utility. The results successfully obtained indicates that reducing the reserved customers' numbers blindly cannot always improve customers' satisfaction, and improving the service quality to all customers is thus necessary, rather than being concerned with only the satisfactions of reserved (VIP) customers. The findings also reveal the dynamic influence mechanism of service intervention time to giving-up customers' numbers, indicating a promising future for applying the methodology framework in improving service quality and customer satisfaction. This work offers new perspectives and valuable demonstration on promoting future strategic researches about service operations management improvement with the integration of customers' behaviors and service operations system simulation though emerging hybrid modeling paradigm.
查看更多>>摘要：The technological leaps of artificial intelligence (AI) and the rise of machine learning have triggered significant progress in a plethora of natural language processing (NLP) and natural language understanding tasks. One of these tasks is argumentation mining which has received significant interest in recent years and is regarded as a key domain for future decision-making systems, behaviour modelling, and natural language understanding problems. Until recently, natural language modelling tasks, such as computational argumentation schemes, were often tested in controlled environments, such as persuasive essays, reducing unexpected behaviours that could occur in real-life settings, like a public debate on social media. Additionally, the growing demand for enhancing the trust and the explainability of the AI services has dictated the design and adoption of modelling schemes to increase the confidence in the outcomes of the AI solutions. This paper attempts to explore modelling argumentation in short text and proposes a novel framework for argumentation detection under the name Abstract Framework for Argumentation Detection (AFAD). Moreover, different proof-of-concept implementations are provided to examine the applicability of the proposed framework to very short text developing a rule-based mechanism and compare the results with data-driven solutions. Eventually, a combination of the deployed methods is applied increasing the correct predictions in the minority class on an imbalanced dataset. The findings suggest that the modelling process provides solid grounds for technical research while the hybrid solutions have the potential to be applied to a wide range of NLP-related tasks offering a deeper understanding of human language and reasoning.
查看更多>>摘要：A new AdDoKu method of photovoltaic (PV) module arrangement for total cross tied (TCT) interconnection pattern has been introduced in this paper. This arrangement is aimed at minimizing the PV array (PVA) losses caused by partial shading conditions and, thereby, increases the power output. Simulations are carried out in an extensive scale to validate the improvement in performance as compared to the TCT connection. Comparisons are also made between the newly proposed AdDoKu pattern and the SuDoKu configuration. The data obtained during this comparison place the proposed scheme at par with the SuDoKu configuration in terms of operational characteristics. The physical location of modules is rearranged without altering any electrical connections. The data obtained from the simulation tests show that the AdDoKu arrangement pattern reduces the number of local peaks in the P-V characteristics and, thereby, making the global maximum power point (GMPP) tracking algorithm much simpler and more accurate. MATLAB/SIMULINK is used for the modelling of all the PVAs considered in this work. Thus, the results establish the performance improvement of the proposed AdDoKu arrangement. Four different types of shading patterns (i.e., short wide, long wide, short narrow, and long narrow) are considered for analyzing the performance of the PVA system based on GMPP, mismatch power loss, % fill factor, % power loss, % power enhancement and energy saving with annual income generation analysis for the proposed topology. The output results shows that the AdDoKu method generates better power than TCT method by 28.76%, 30.15%, 11.66% and 0.34% in short wide, long wide, short narrow and long narrow shading pattern, respectively. Furthermore, the proposed AdDoKu method shows an improvement of 1.94% as compared to SuDoKu method under different shading conditions.
查看更多>>摘要：To face the booming passenger flow volumes, the optimization of pedestrian facilities plans and the associated staff assignment plans in the transfer station, including adjusting pedestrian flow routes and adopting different numbers of machines, is an efficient measure to promote transfer connection and avoid passengers congestion in order to enhance the service level. We propose an integrated optimization model based on the simulation model and machine learning method aiming to provide optimal pedestrian facilities plans and the corresponding staff assignment simultaneously. The pedestrian facilities plans ensure the service quality requirement based on three performance indicators, i.e., transfer capacity, transfer average time and level-of-service. The framework, which contains the transfer station simulation model and the Random Forest for obtaining the values of performance indicators, is developed. Random Forest, a machine learning method, trains and learns the samples of indicators and scenario attributes outputted from the simulation model to fit the approximation functions that can evaluate the performance indicators under the given scenarios. As the application, the optimization model uses the approximation functions to obtain performance indicators so that select out the available pedestrian facilities plans for time-varying passenger demands. With the aims of minimizing employment cost and preventing excessive fatigue, the optimization model selects the most suitable pedestrian facilities plans and obtains the corresponding staff assignment plan with the consideration of the workload fairness and rest time under four types of consecutive working time constraints. The experiments of the Xipu station demonstrate that the proposed integrated optimization model can return reasonable pedestrian facilities plans and staff assignment plans for each period in a day. The constraints of consecutive working time can acquire good compromising of reducing labor cost and avoiding overwork, so that the staff assignment plans are acceptable for both station managers and employees.
查看更多>>摘要：Development of control algorithms for enhancing performance in safety-critical systems such as the Autonomous Emergency Braking (AEB) system is an important issue in the emerging field of automated electric vehicles. In this study, we model a safety distance-based hierarchical AEB control system constituted of a high-level Rule-Based Supervisory control module, an intermediate-level switching algorithm and a low-level control module. The Rule Based supervisor determines the required deceleration command that is fed to the low-level control module via the switching algorithm. In the low-level, two wheel slip control algorithms were developed, a Robust Sliding Mode control algorithm with an Artificial Neural Network (ANN) for nonlinear parameter estimation and a Gain-Scheduled Linear Quadratic Regulator. For the needs of this control design, a non-linear dynamic vehicle model was implemented whereas a constant tire-road friction coefficient was considered. The proposed control system was validated in Simulink, assuming a straight-line braking maneuver on a flat dry road. The simulation results demonstrated satisfactory emergency braking performance with full collision avoidance in both proposed control system combinations.
Mehmet Mert IlmanSahin YavuzHira KaragulleAli Uysal...
25页
查看更多>>摘要：This paper introduces a novel hybrid vibration control of an industrial flexible link manipulator (FLM) with carbon-fiber reinforced polymer (CFRP) composite material. The hybrid controller consists of a closed-loop direct strain feedback controller (DSFC) and a new open-loop input shaper (OLIS), and its performance is examined and compared by numerical simulation and experiment. Vibration attenuation is also verified by a spare sensor, an accelerometer. Thanks to the design of an analog isolator circuit, data acquisition of the real-time experiments are carried out noise-free from the servo driver. Reduced-order-model (ROM) is generated for the composite link with three different loading conditions and model for servo system is derived via system identification. Impulse force test hammer (IFTH) equipment is used to excite the system with force input and the results of the experiments/simulations confirm that the controller is robust against the impulse disturbances. As a result, a decrease of 88 ± 6.9 percent in vibration amplitudes and a decrease of 90 ± 7 percent in vibration duration confirms that the proposed controller is effective for vibration control of the flexible arm even under disturbance. Controller performance also enabled the FLM to work with payloads 16 times of its own weight.
查看更多>>摘要：In this study, simulation approaches for the reliability estimation of a multi-state network (MSN) are developed; the approaches obviate the reliance on known capacity probability distributions. Two procedures are proposed to randomly generate the system (capacity) state, which is further compared with the lower bounds of demand, for estimating system reliability. In particular, the simulation approaches consider the time-varying reliability function for components in the MSN. Two examples, i.e., with and without considering time-varying reliability functions, are investigated to determine the effectiveness (accuracy, convergence, and variance) and efficiency of the proposed procedures. Additionally, a practical case of the National Science Foundation Network is investigated to validate the scalability and applicability of the proposed procedures. Experimental results show that both proposed procedures are effective and efficient for estimating the system reliability of an MSN.
Jairo Romero HuertaRanyere Sousa SilvaGiorgio De TomiAnna Luiza Marques Ayres da Silva...
11页
查看更多>>摘要：Underground mining is a highly competitive industry that is often singularly focused on material production as a key performance indicator. However, maximizing material production does not always maximize industry profitability. Equipment utilization in support of underground mining operations is a significant factor that affects both material production rates and industry profitability. Specifically, uncertainties related to equipment fleet acquisition, operations, and maintenance can challenge the predictability of profit margins based on material production. With a focus on maximizing industry profitability, a methodology to support operational decision-making in underground mining based on dynamic simulation was developed. First, data related to underground mining operations from a medium-sized mine in southern Peru were collected and processed. Based on them, relevant operational indicators were selected and defined for inclusion in the simulation model. Next, the simulation model was formulated and applied to four scenarios with distinctive equipment fleets and associated material production rates. Simulation results indicated that the scenario with the lowest equipment fleet costs and the lowest associated material production rate (base case) had a profitability margin that was nearly 13 times higher than the scenario with the highest equipment fleet costs and the highest associated material production rate. This dynamic simulation methodology, which was successfully demonstrated in this study, can be broadly applied to support operational decision-making in underground mining to maximize industry profitability based on a wider and holistic array of factors beyond material production.
查看更多>>摘要：Making clear the metallurgical transport behaviors during direct chill (DC) casting is of great significance to control the solidified structure and reduce the casting defects, but it still remains difficult because the high temperature and the opaque of melt make them hard to observe and detect directly. Numerical modelling provides a way to reveal the physical fields involved during DC casting. The present research used a two-dimensional axisymmetric model which coupled the heat transfer, fluid flow, solidification and species transfer in order to study the effect of different magnetic fields (harmonic magnetic field (HMF), pulsed magnetic field (PMF) and out-of-phase pulsed magnetic field (OPMF)) on the flow pattern, solidification characteristic and solute distribution in the electromagnetic DC casting of magnesium alloy billet having the diameter of 500 mm. The results show that when the magnetic field is on, there are the desirable variations including the homogenous temperature profile and faster solidification rate, as well as the alleviated central negative segregation and the reduced whole segregation level on the cross section of the billet. For the different magnetic fields, OPMF shows the strongest stirring effect, leading to the highest solidification rate at the initial solidification stage. Furthermore, the central negative segregation and the segregation level along the cross section are alleviated most as compared to other magnetic fields.
查看更多>>摘要：The determination of the traffic carrying capacity of road network is of great significance to the scale control of urban vehicles and the control of transportation systems. Existing related works have proposed calculation models and solutions for different application environments. However, these studies have problems such as diversification of capacity definitions, incorrect analysis of factors, and excessive model constraints. In response to these problems, this paper aims to determine the connotation of traffic carrying capacity and put forward a novel calculation model for carrying capacity. Firstly, in-depth analysis found that the key factor affecting the number of vehicles is the service level of road network. Then, the connotation of traffic carrying capacity is clarified, that is, the maximum number of vehicles travelling simultaneously in the road network under a certain service level. Second, based on the definition of carrying capacity and the characterization relationship between average travel speed and service level, the calculation of carrying capacity is transformed into the establishment of relationship between average travel speed and maximum number of vehicles. The construction of this relationship includes five steps, which are proposal of equivalent lane, division of traffic basic units, construction of basic unit travel time model, calculation of basic unit carrying capacity, and calculation of road network carrying capacity. In particular, the proposal of equivalent lane achieves the conversion from intermittent flow to continuous flow, and the abstraction of traffic flow further unifies the operation process into a unified mode of free travel and stacked release. Finally, the VISSIM simulation software is employed to verify and evaluate the proposed models. The results show that the average relative error between the simulated data and the calculated data obtained from the model is -0.46%. Findings from this study will provide an innovative idea for the evaluation of urban road network performance. It can also provide a basis for urban managers to improve service level of road network and optimize the design of transportation system.
NSTL
Elsevier